I’ve been saying for several years that the best person to manage our way out of the current economic doldrums would be George Balanchine (if only he weren’t dead) – because this will need to be the most carefully choreographed dance of monetary and fiscal policy in all history. The Federal Reserve will at some point need to raise interest rates that currently are on the floor (“at the zero bound,” in econ speak) and draw down a balance sheet that is orders of magnitude greater than its normal size, to head off inflation – all in a shaky economy nestled in a shaky world economy. Meanwhile, the fiscal policymakers will have to head off a mounting debt by slashing a far-oversized budget deficit, which is driven by complex structural problems with their own powerful political self-defense mechanisms – all the while avoiding crunching that same vulnerable economy, and somehow acting in harmony with the aforementioned independent Federal Reserve. It is a situation only an academic economist could love: It offers plenty of ivy-covered reward for writing theoretical papers which will never be tested in practice, and so have no real-world consequences.
And that is the good news. A recent more-practical (but still plenty wonkish) paper by two Federal Reserve economists, Christopher J. Erceg and Andrew T. Levin, explains that today’s labor market is not only painful, but also puzzling to policymakers. It identifies yet another unprecedented challenge that is layered upon all the others to make the path back to Normal, wherever that is (unfortunately not the town that is readily visible on the map of Illinois), even more tortuous.
The Erceg and Levin paper already has gotten plenty of press (for example, see a New York Times reference here), but is worth your attention if you have not seen a close discussion.
Here is the problem, in a nutshell: In a “normal” economy, adults tend to work or to look for work – “labor force participation” in the technical jargon – in quite stable patterns. Younger people go to school or to work, older people work or retire, and women work or bear and raise children, in relatively predictable proportions. There are long-term trends in each of these decisions, but they move fairly slowly. And though this behavior changes in economic downturns, it does not change very much.
The last six years, however, have been distinctly non-normal. Starting in late 2007, the economy endured a much-oversized recession. Into 2009, Erceg and Levin estimate that labor force participation was pretty much on the track that would be predicted on the basis of history. But in 2009, participation began to drop below a reasonably predicted path. And tellingly, participation dropped the most where the recession – by then, by far the worst in post-World War II U.S. history – hit the hardest. If there had been some general shift in attitudes, perhaps driven by national policy, one would have expected the change to have been uniform from sea to shining sea. It wasn’t. The poster children were Arizona, California, Florida and Nevada. Think real-estate meltdown, and you’ll get the idea.
And even further from normal, as the economy has recovered in the tentative way that we have seen so far, the unemployment rate has come back – down about halfway from its peak to where it was when the recession began – but the labor force participation rate has not. It remains virtually at its trough. In other words, the people who left the workforce have not rejoined (all of this on average and with flows both ways, of course); rather, some of the people who lost their jobs but kept on looking for work have found it.
Based on that finding, Erceg and Levin pose the problem – and coming from the staff of the Federal Reserve, it is appropriately centered on monetary policy: The Fed normally makes its decisions with one eye on the unemployment rate. But in this abnormal time, will the unemployment rate be a predictable and accurate indicator of the state of the labor market?
The Fed’s mission is to achieve maximum employment consistent with low and stable inflation. With about two-thirds of the cost of production in the United States being the cost of labor, price stability requires maintaining wage increases relatively close to the rate of growth of productivity. (In recent years, wages have lagged productivity – probably because demand has been so weak, and joblessness so great.) We all understand that wages follow the tightness in the labor market; when available workers are relatively few, employers must bid higher wages to get the workers they need.
The Fed’s program at this point is to maintain its stimulative posture at least until the unemployment rate falls to 6.5 percent (in other words, by one more percentage point), provided that inflation remains controlled. But might the large reserve of former workers who have left the labor force decide to re-enter after the unemployment rate has fallen enough to prompt the Fed to tighten? Could the 6.5 percent unemployment rate prove to be a “false floor” on the zone where Fed policy should be stimulative? Tightening too soon would at least sacrifice potential non-inflationary economic growth, and could even bring back recession. But holding the accelerator to the floor too long would risk a costly and persistent bout with inflation.
This situation is perilous precisely because the economy is off of the historical charts. Erceg and Levin provide estimates of potential outcomes, but they are appropriately modest about their precision.
A potentially highly useful part of Erceg and Levin’s paper is an exploration of subgroups of the potential workforce by age, to see to the extent possible who has chosen to leave the workforce, and therefore potentially how likely they would be to re-enter. What they find is that a significant share of the workforce withdrawals has come from younger workers, many of whom apparently have scanned a forbidding job market and determined that they might as well stay in school. Interestingly enough, the oldest workers (65 and over) appear to have stayed in the job market beyond expectations. Their choices seem to follow the anecdote that some older persons lost so much of the value of their portfolios and the equity in their homes that they could not afford to retire. Such current older workers might be expected to leave the labor force once they see their way clear to a reasonable retirement. The most questionable group is the so-called “prime-aged” workers (25 to 54, and 55 to 64 years old), who have withdrawn from the workforce in greater-than-anticipated numbers. Of those, some in the 25 to 54 year-old group have re-entered school; a significant number from 55 to 64 years old have claimed disability coverage under Social Security. The former are of course likely to re-enter the labor force when times are better, perhaps to earn a return on their investments in schooling; the latter are highly unlikely to rejoin the work force.
So if you buy the Erceg and Levin reasoning and modeling, you would counsel the Federal Reserve to maintain a stimulative posture for longer than the usual unemployment-rate indicator would suggest. The reason would be that at some point, the discouraged workers are likely to begin to rejoin the labor force and search for work. Those job seekers will reduce the upward pressure on wages, and facilitate further economic growth with less-than-predicted risk of inflation. But this is a dangerous game; it is as though you were informed while driving down a superhighway that your speedometer overstates your speed – but by an unknown margin. You may press harder on the accelerator, but you do so at increasing peril.
I can add one potentially relevant piece of data to the Erceg and Levin story – which was presented in a chart in an earlier blog and in a different context. As noted earlier, Erceg and Levin found that withdrawals from the labor force occurred disproportionately in states that were hit hardest by the housing bust. It turns out that the drop in employment in the segments of the construction industry (residential building and land subdivision) most directly related to housing were enormous. Jobs in residential building fell by almost 45 percent from their level of early 2006 (which admittedly was probably beyond the sustainable); jobs in land subdivision dropped by a staggering 55 percent. Neither category has seen much of a recovery thus far. (See chart below.)
We economists speak blithely about displaced workers migrating instantaneously and costlessly into their next-best uses. But such migration is never instantaneous or costless, and sometimes it isn’t possible at all. And it can be especially out-of-bounds for home-building workers. First, the people who drive Erceg and Levin’s numbers live in the hardest-hit parts of the country – where the home-building crunch was so strong that it affected the rest of the economy to the greatest degree. That means that even if those construction workers have the aptitude to become, say, computer network engineers, they live where the number of such jobs is likely to be the least. And by definition, because they come from down housing markets, they are the most likely to be underwater on their mortgages and to have the hardest times selling their homes if they should be willing and able to move on every other score.
In other words, we might find that the resolution of the Erceg and Levin conundrum will have to wait on a housing recovery. And with apologies to Tip O’Neill and his descendents, all housing is local. We have had regional economic phenomena before – oil booms and oil gluts, bi-coastal economies (up and down), worldwide food-price fluctuations, and everything else. But this downturn has been the biggest, wildest ride in my lifetime, at least, and it is likely to remain so for some time. There will be local and regional disparities in the economy – which pose their own problems for national policymaking.
So thanks to Erceg and Levin for a clearer picture of a dauntingly bad scene. And to my friends on the Federal Open Market Committee: You’ve earned your combat pay, at least for this quarter.